Zobrazeno 1 - 10
of 262
pro vyhledávání: '"Ibrahim, Adam"'
Autor:
Ibrahim, Adam
Les modèles d'apprentissage automatique (ML) atteignent des performances remarquables sur les tâches pour lesquelles ils sont entraînés. Cependant, ils sont souvent sensibles aux changements dans la distribution des données, ce qui peut nuir à
Autor:
Schaeffer, Rylan, Schoelkopf, Hailey, Miranda, Brando, Mukobi, Gabriel, Madan, Varun, Ibrahim, Adam, Bradley, Herbie, Biderman, Stella, Koyejo, Sanmi
Predictable behavior from scaling advanced AI systems is an extremely desirable property. Although a well-established literature exists on how pretraining performance scales, the literature on how particular downstream capabilities scale is significa
Externí odkaz:
http://arxiv.org/abs/2406.04391
Autor:
Tokpanov, Yury, Millidge, Beren, Glorioso, Paolo, Pilault, Jonathan, Ibrahim, Adam, Whittington, James, Anthony, Quentin
The size of large language models (LLMs) has scaled dramatically in recent years and their computational and data requirements have surged correspondingly. State-of-the-art language models, even at relatively smaller sizes, typically require training
Externí odkaz:
http://arxiv.org/abs/2406.01981
Autor:
Glorioso, Paolo, Anthony, Quentin, Tokpanov, Yury, Whittington, James, Pilault, Jonathan, Ibrahim, Adam, Millidge, Beren
In this technical report, we present Zamba, a novel 7B SSM-transformer hybrid model which achieves competitive performance against leading open-weight models at a comparable scale. Zamba is trained on 1T tokens from openly available datasets and is t
Externí odkaz:
http://arxiv.org/abs/2405.16712
Autor:
Ibrahim, Adam, Thérien, Benjamin, Gupta, Kshitij, Richter, Mats L., Anthony, Quentin, Lesort, Timothée, Belilovsky, Eugene, Rish, Irina
Large language models (LLMs) are routinely pre-trained on billions of tokens, only to start the process over again once new data becomes available. A much more efficient solution is to continually pre-train these models, saving significant compute co
Externí odkaz:
http://arxiv.org/abs/2403.08763
Autor:
Ibrahim, Adam A.
Introduction: In 2012, approximately 15 % of human cancer cases were attributable to infectious agents worldwide. Fibropapillomatosis is a marine turtle disease characterized by growth of tumors on the shell, skin, eyes, oral cavity and/or viscera of
Externí odkaz:
http://urn.kb.se/resolve?urn=urn:nbn:se:oru:diva-109006
Autor:
Ibrahim, Adam, Ketsela, Isak
The purpose of this study is to explain the change that occurs in the capital structure when a company goes public and to review whether there are any possible factors and theories who can explain the change. The data has been collected using a quant
Externí odkaz:
http://urn.kb.se/resolve?urn=urn:nbn:se:sh:diva-50670
Autor:
Gupta, Kshitij, Thérien, Benjamin, Ibrahim, Adam, Richter, Mats L., Anthony, Quentin, Belilovsky, Eugene, Rish, Irina, Lesort, Timothée
Large language models (LLMs) are routinely pre-trained on billions of tokens, only to restart the process over again once new data becomes available. A much cheaper and more efficient solution would be to enable the continual pre-training of these mo
Externí odkaz:
http://arxiv.org/abs/2308.04014
Autor:
Ibrahim, Adam
Purpose The present review aimed to evaluate the difference of dental implant failure rates and marginal bone loss (MBL) between implants inserted in fresh extraction sockets or healed sites. Materials and methods Electronic search was undertaken in
Externí odkaz:
http://urn.kb.se/resolve?urn=urn:nbn:se:mau:diva-51774
Autor:
Ibrahim, Adam, Guille-Escuret, Charles, Mitliagkas, Ioannis, Rish, Irina, Krueger, David, Bashivan, Pouya
Adversarial robustness continues to be a major challenge for deep learning. A core issue is that robustness to one type of attack often fails to transfer to other attacks. While prior work establishes a theoretical trade-off in robustness against dif
Externí odkaz:
http://arxiv.org/abs/2210.03150